ROS-based Navigation Software For Handling Robots

Navigation technology for handling robots based on the storage environment has now become a key focus of research in the smart storage industry.This paper relies on the project "Design of a navigation control system for stacking robots",in which an ordinary forklift truck is transformed into a handling robot with autonomous navigation capabilities.The paper takes the establishment of an indoor map of the handling robot and the path planning technology based on the map model as the starting point,and presents a scheme for the design of the navigation system of the handling robot.The main parts of the research in this paper include the following:Firstly,the thesis takes a branded electric forklift truck as the carrier from the system and user requirements,and use the self-developed intelligent controller and embedded micro industrial control card and sensors such as LIDAR to form a hardware development platform for the handling robot.The software platform design of the handling robot navigation system is combined with the system requirements and the hardware platform,and the software design and visualisation of the modules are presented.Then,the laser SLAM-based method for building maps of the storage environment is studied and implemented.The principles and formulas of the Gmapping SLAM,Hector SLAM and Cartographer SLAM methods are analysed and derived.The three algorithms were tested in a simulation environment and a physical environment using cosine similarity and structural similarity as evaluation indicators for map construction accuracy.The experimental results show that the Cartographer SLAM method builds maps with higher accuracy and is more suitable as a map building method for the manned storage environment of the handling machine in this paper.An improved A* algorithm based on the minimum safe distance between convex corridors and obstacles is then proposed for the path safety needs of handling robots navigating in a warehouse environment.The distance of the path from the nearest obstacle is used as the safety evaluation index and the running time of the navigation process is used as the efficiency evaluation index to compare the traditional A* method with the improved A* method in the constructed map model for path planning experiments.The experimental results show that the improved A* algorithm has significant advantages over the traditional A* algorithm in terms of both path safety and navigation efficiency.Finally,the authors of this paper have completed a software implementation based on the above design and carried out practical engineering tests.The test results show that the handling robot system in this paper has the ability to visualise the map construction at the application level,and the map construction process and the interaction of the map construction process can be done at the interface side;the robot has the ability to visualise the navigation at the application level,and the paths planned by the improved A* algorithm and the positional data during the robot navigation can be fully expressed at the interface side;the robot navigation accuracy meets the needs of enterprises;the system functions are stable and practical.The system is stable and practical.